Get your web usability testing right with these tips from Danny Bluestone, MD of Cyber-Duck
The most successful user-centred websites are built on usability testing, as no amount of web production experience parallels feedback from real users, and insights gained from Google Analytics and other quantitative testing tools (such as A/B testing) cannot be used alone. Perceived expense and tight timelines corner many companies into delaying testing until just before launch, severely limiting the capacity to action insights gathered. When done correctly, usability testing is carried out throughout the project lifecycle, to guide and challenge design decisions. These top 10 tips will help you to avoid potential pitfalls and ensure you get the most out of usability testing:
1. Limit the numbers tested
Studies have revealed how employing large numbers of users for testing is a waste of resources. The first five participants will uncover around 80 per cent of all usability issues, so running smaller tests minimises unnecessary repetition of results. Similarly, Microsoft’s experience designing the Office suite reveals that if 20 per cent of the most noted bugs are dealt with, 80 per cent of the issues will be eliminated.
2. Test early, and often
Targeted tests should be run frequently throughout the development process. They should be incorporated into early research stages, with previous websites being tested (as a ‘benchmark’ for comparison) in redesign projects; this should continue right through to user acceptance testing (UAT) before launch. Repeated testing enables the user feedback to be addressed and subsequent tests to be run on the new and improved design. These subsequent tests often reveal more significant usability insights that were masked by surface-level problems in earlier rounds.
3. Recruit representative users
Don’t underestimate the importance of testing on the right participants; testing on the ‘wrong’ people can heavily skew the results. The most valuable, representative participants are real samples of your different types of customers and prospects, as each will interact with your website in different ways. These can be gathered from your CRM or using specialist recruiters.
4. Set tasks based on goals
The tasks set should be inspired by the key goals of the website, essential for its success: for example, the ease with which users can find and buy products from an ecommerce store. Instead of providing instructions, tasks should be made realistic and be given context, presenting them as ‘scenarios’; for instance, telling the user they need to contact the company urgently, to enquire about a missing delivery.
5. Encourage natural behaviour
The most accurate results are secured through encouraging users to act as naturally as they can in the artificial environment. Moderators can help to put participants at ease, they should encourage the user to behave as if they were on their own, reminding them that data generated is confidential.
6. Collect measurable data
The subtleties of a user’s behaviour during testing can be best captured using a variety of technologies that reveal behaviours the user may not even be aware of, and allow the moderator to review the test in retrospect. Tools like Morae record mouse movements, as well as video and audio feeds. Eye tracking technology from Tobii uncovers issues particularly effectively, as it monitors where the user’s eyes are fixating on the website in relation to mouse or finger movement, as well as the path and speed of eye movement, facial expressions, and length of time on each area.
7. Include subjective insights
Quantifiable data should be combined with more subjective, emotive insights to achieve the best results and a holistic picture of how the website is performing for users. Techniques like recording the user’s spontaneous verbalisations, and retrospective think-aloud (RTA) are the most effective. This sheds light on how the users emotionally connect with the website.
8. Choose the environment
The best environment often depends on the specific testing scenario. Formal testing is conducted in a lab or office. It is more controlled, often having facilities such as two-way mirror testing rooms with audio/visual equipment and enables users to be tested individually and talk about their experiences freely. Informal testing is far more spontaneous, and thereby compatible with any venue. It generally involves persuading users to take part with the study in situ, and moderators observe their performance. While it lacks the control of the formal environment, it does have the benefit of contextual relevancy.
9. Collate actionable insights
Designers and developers need to receive actionable insights as an output of the usability testing, so that the feedback collected can inform design decisions and influence production. The results should be collated into a concise report, listing a set of required changes that result from critical analysis of the findings.
10. Test, test and test again
Not all of these tips can always be followed, but the biggest mistake you can make is to use that fact as an excuse to avoid usability testing altogether; it is far better to keep testing, while taking the limitations and constraints into account in your analysis.
